AI IN NPD Can AI compete with the crowd in generating ideas?
Keywords
Loading...
Authors
Issue Date
2023-09-28
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
This study explores the idea generation capabilities for new potato products from AI (represented by
a Transformer Based Language model called ChatGPT) in comparison to the ideas collected via
crowdsourcing. ChatGPT model GPT-4 was given the same questions and information as the crowd
was given. In total 206 ideas, 103 ideas from both sources, were rated by the manager innovation and
a product developer from Company X on four dimensions. The dimensions are novelty, feasibility,
relevancy and specificity, together forming idea quality. Results from a MANCOVA and a Mann-
Whitney-U test showed that the crowd scores significantly higher on novelty and specificity than AI
does. The source of the idea did not have a statistically significant effect on the feasibility of the ideas,
so the feasibility is not dependent of the source of the idea. This indicates that the implementability
of the ideas from the crowd and AI are almost equal. The influence of the source of the idea on
relevancy could not be measured do to negative inter-rater reliability. The results need to be
interpreted with caution due to non-normality of the data.
Description
Citation
Supervisor
Faculty
Faculteit der Managementwetenschappen